government.* Those who start with a concern for efficiency and cost are likely, therefore, to emphasize the importance of restructuring existing governments into larger units that may be more capable of solving problems at lower cost. In contrast, those who start with a concern for democratic participation are likely to give attention to how to delegate authority from central, state, and municipal governments to local, neighborhood bodies. The conflict between these two approaches was well stated by Dahl in his presidential address to the American Political Science Association in 1967. The larger and more inclusive a unit, the more its government can regulate aspects of the environment that its citizens want to regulate, from air and water pollution and racial injustices to the dissemination of nuclear weapons. Yet the larger and more inclusive a unit with a representative government, and the more complex its tasks, the more participation must be reduced for most people to the single act of voting in an election. Conversely, the smaller the unit, the greater the opportunity for citizens to participate in the decisions of their government, yet the less of the environment they can control. Thus, for most citizens, participation in very large units becomes minimal and in very small units it becomes trivial. At the extremes, citizens may participate in a vast range of complex and crucial decisions by the single act of casting a ballot; or else they have almost unlimited opportunities to participate in decisions over matters of no importance. At the one extreme, then, the people vote but they do not rule; at the other, they rule—but they have nothing to rule over. (47; see also 48) Without attempting to reconcile these conflicting values, Dahl concludes that the optimal city size is in the broad range from about 50,000 to 200,000, but this number seems to reflect the fact that he lives and works and is obviously happy in New Haven, which falls within this range. The size of governmental units and the distribution of power from one governmental unit to another is perhaps one of the most significant and correspondingly difficult problems in the political development of both modern and developing societies. In modern societies it is a problem because technological changes have made existing units too small to cope with their problems; and in developing societies it is a problem because populations are often so large, dense, and ethnically divided that highly centralized government is often not acceptable, and provincial and local governments may be *There can be diseconomies of scale, too, since beyond a certain size the costs per unit may begin to rise. Sundquist (46, p. 92) points out that for some large metropolitan areas the costs of water supply, sewage, and solid waste disposal may rise for the simple reason that the water and waste must be carried over longer distances. Moreover, many pollution problems are made worse by high population—air pollution, for example, as a result of the dense concentration of automobiles. The cconomies-of-scale argument, therefore, cuts both ways. less important.